Strategies to reduce attrition in randomised trials

Background

Attrition from randomised trials can introduce bias and reduce study power affecting the generalisability, validity, and reliability of results [1]. Many strategies are used by trialists to reduce attrition, including motivating and engaging participants and sites to optimise data return or compliance to follow-up procedures [2].

Objective

To quantify the effect of strategies to reduce attrition from randomised trials in any healthcare setting.

Methods

Included studies were randomised evaluations of strategies to reduce attrition embedded within randomised trials from all disease areas and settings. The following sources were searched for eligible studies [3]: MEDLINE (1950 to present), EMBASE (1980 to present), PsycINFO (1806 to present), DARE (most recent issue), CENTRAL (most recent issue), CINAHL (1981 to present), C2-SPECTR (most recent date), and ERIC (1966- present), Cochrane Methodology Register, Current Controlled Trials meta Register, WHO trials platform, Society for Clinical Trials (SCT) conference proceedings (1980-2010), and publication reference lists. A survey of all UK clinical trials units (CTU) was also conducted to identify studies.

Two authors reviewed potentially eligible titles and abstracts. Data extracted were checked by two authors. Study investigators were contacted for missing data. Risk of bias was assessed using the Cochrane risk of bias tool. Data were entered into RevMan5 and pooled using the fixed effect model. Heterogeneity was explored to determine whether some types of strategies to reduce attrition were more effective than others. The analyses focused on the primary endpoint of attrition.

Results

From 19,281 abstracts 31 unique RCTs were identified from the following sources: MEDLINE, CENTRAL, CINAHL n=9; SCT abstracts 1980-2010 n=4; reference lists of relevant reviews n=7; and of included trials n=8 (7 duplicates); word of mouth n=4; and CTUs survey n=6. Six types of strategies to reduce attrition were identified: a) communication i.e. email, letters signed by different study personnel, type of post, and delivery method; b) questionnaire length i.e. short versus long; c) incentives i.e. monetary incentives, offers of monetary incentives or vouchers, and gifts; d) case management i.e. trial assistants assigned to manage participant follow-up; e) behavioural e.g. workshops giving participants information about goal setting; and f) methodological interventions e.g. blinded versus open trials. Final results of the review will be presented.